76 J. C. ECCLES 



e.p.s.p. However, when the rising phase of the i.p.s.p. precedes the onset of 

 the e.p.s.p. (Fig. 4b-f), there is no potentiation of the later decaying phase 

 of the i.p.s.p., as is well illustrated in the subtracted records beginning at 

 —4 and —2 msec in Fig. 4n. This absence of effect would be expected if the 

 inhibitory current ceased within 2 msec of its onset as shown in Fig. 2a. 

 On the other hand there is potentiation of the i.p.s.p. when it is generated at 

 any stage of the declining phase of the e.p.s.p., and, as would be expected 

 from Fig. 2c, this potentiation closely follows the time course of the e.p.s.p. 

 (Figs. 4n, o). 



Although it has been shown that, by the latency test, the i.p.s.p. is precisely 

 fitted to be the initiator of inhibitory action on reflex discharges (Araki et al., 

 1960; Eccles, 1961), it is necessary to see whether this good agreement holds 

 for the whole time course of the inhibitory action. 



The time course of inhibitory action can be ascertained by observing the 

 depression of a testing monosynaptic reflex discharge at varying times after 

 the conditioning inhibitory volley. In the original descriptions the inhibitory 

 curve so obtained reached a maximum with a volley interval of less than 

 1 msec, and then declined along an approximately exponential curve with a 

 time constant of about 4 msec (Lloyd, 1946; Laporte and Lloyd, 1952). 

 However, in more recent investigations where special precautions were taken 

 to have the inhibition produced by virtually pure la volleys, an initial very 

 rapid phase of decay has declined on to a slowly decaying residuum, as illus- 

 trated in Fig. 5a, c (Bradley et al., 1953; Brooks et al., 1957; Araki et al., 

 1960). Usually inhibitory curves with this double composition were also 

 observed by Jack et al. (1959, and personal communication), but occasionally 

 they observed only the brief initial phase. It was suggested that in such 

 circumstances the spinal cord was in particularly good condition and that the 

 membrane potential of the motoneurons was then as high as the equilibrium 

 potential for the i.p.s.p., i.e. that Er = -S'i.p.s.p. 



Before describing further investigations into the time course of inhibitory 

 synaptic action, it is desirable to show how the double composition of in- 

 hibitory curves is related to the mode of operation of inhibitory synapses. 

 It was first suggested by Coombs et al. (1955c) that there was a brief intense 

 inhibitory effect superimposed on a more prolonged action that had a time 

 course corresponding to the i.p.s.p., as is illustrated in Fig. 5b, c. It was 

 further suggested that the initial intense inhibition was directly due to the 

 current flow generated by the activated inhibitory synapses, i.e. to the current 

 shown in Fig. 2a. Such currents would directly antagonize the depolarizing 

 currents generated by activated excitatory synapses, their effectiveness being 

 enhanced if superimposed on a depolarization (the e.p.s.p.) already produced 

 by the excitatory synapses. However, the inhibitory current has a duration of 

 no more than 2 msec, and excitatory synapses activated after this time would 

 be antagonized only by the residuum of hyparpolarization. In contrast, at 



